Indi an Journal of Radio & Space Physics Vol. 28, October 1999, pp. 226-23 1

Long period atmospheric oscillations in the troposphere and lower and middle stratosphere over Hyderabad

H Aleern Basha l, M N Joshi2, B Bapiraju l

, S Srinivasan2, J V Subba Rao2 & Gopa Dutta l

INawab Shah Alam Khan Centre for Post-Graduate Studies & Research (Anwarul-uloom College), New Mallepally, Hydcrabad 500 00 I

2Tata Institute of Fundamcntal Research, Balloon Facility, Hyderabad 500 062

Received 24 August 1998; revised 12 August 1999; accepted 6 September 1999

The monthly mean values of zonal and meridional winds for seventeen years over a tropical station, Hyderabad (17.2° N;78.2° E) have ~cen used to study the characteristics of long period atmospheric oscillations, the QBO, AO, SAO and TAO, in the altitude range 7 - 32 km. Since the period of the data includes two solar maxima and one solar min imum the solar activity effects on various atmospheric parameters have been investigated. The annual oscillation in zonal wi~d is found to be strongly correlated with solar activity, whereas the prevailing zonal wind shows poor correlation. The correlation of solar activity with meridional wind parameters is also found to be very poor. The period of QBO and its heights of maxima are found to be affected by solar activity variation, whereas its amplitude is not.

1 Introduction

It is well established that long period oscillations I ike the quasi-biennial osci llation (QBO), annual oscil lation (AO), semi-annual oscillation (SAO) and ter-annual oscillation (TAO) dominate the tropica l middle atmosphere l

-4. As the atmosphere of the earth is mostly solar dependent, various phenomena in the middle atmosphere are found to be affected by solar variabil ity. The QBO is reported to show variability from cycle to cycle in its period and a·mplitude5

-7

• The coupli ng of the wind system with the 1 I - yr solar cycle has been subjected to various investigations 8- 10.

Some of the studies show contradictory results making the subject still controversial. Most of these reports are from mid-latitude stations and results from tropical latitudes are very sparse. In the present study, we have made a detailed analysis of the long period oscillations and the solar activity variation of wind parameters considering data from a tropical station, Hyderabad (17.2"N,78.2°E).

2 Data and method of analysis

Extensive w ind data have been collected by the National Scientific Bal loon Facility of the Tata Inst itute of Fundamental Research (TIFR), Hyderabad, jointly with the India Meteorological Department (IMD), Hyderabad, for the past two decades by using a special type of rubber and polyethylene balloon capable of reaching up (0 an

altitude of 10 km. The data collected over 17 years (1977-1993) in the a ltitude range of 7-32 km have been utilized for the present study. Wind velocity components have been derived by using raw data consisting of range, elevation and azimuth of the balloon at 1 mVi interval. Generally, there are a few flights every month. Altitude profiles of zonal and meridional winds given by these flights have been averaged to obtain monthly mean profiles. The TJFR restricts balloon flights during monsoon season. Data for such missing months between 1984 and 1993 , over Hyderabad, have been obtained fro m India Meteorological Department, Pune. Very few data gaps which were still left over have been filled by spline interpolation to obtain a continuous data set of 204 months. Month ly break-ups of balloon ascents are shown in Table I.

The continuous data sets of zonal and meridional winds, so obtained, have been Fourier transformed to identify the dominant long period oscillations in the troposphere and lower and middle stratosphere. To obtai n quantitative information about these oscil lations, the data sets have been ubjected to harmonic analysis with least square fitti ng, assuming that the oscillations consist of AO, SAO, TAO and QBO. The periods of AO, SAO and TAO have been taken as 12, 6 and 4 months, respect ively, whereas the period of QBO has been varied between 22 and 34 months to get the best fit.

To study the so lar activity de endence of the

ALEEM BASHA et al. : A TMO,')PHERIC OSCIlLA n ONS IN TROPOSPHERE & STRATOSPHERE 227

Table I - Monthly break-up of balloon ascents during 1977 - 1993

Alt.

10km

15km

20km

25km

30km

Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec.

94

93

93

85

81

115

110

93

76

56

112 112 114 181 169 186 176 155 58 81

80

79

67

60

107

105

98

85

112

101

97

94

113

112

10 1

85

168

137

98

52

amplitude of these long period oscillations, the long zonal and meridional wind data sets have been divided into three periods ( 1978-- 1982), (1983- 1987) and (1988-1993), with average sunspot numbers as 132, 35 and 129, respectively, i.e. covering two high activity periods an~ one low activity period which will henceforth be referred to as HA 1, LA and HA2. Hannonic analysis was then carried out for each group separately to find the amplitudes of these oscillations. We have also calculated cross-correlation coefficients of the yearly averaged sunspot numbers and F I07 cm solar flux with the prevailing wind and the amplitude of AO.

3 Results and discussion

Figure 1 shows FFT spectra of the zonal winds for two particular heights. It is found that AO dominates the troposphere, having maximum amplitude near 12-14 km, whereas QBO starts becoming prominent In

the lower stratosphere, max imizing near 29-30 km .

3.1 Annual oscillation

The annual oscillation (AO) dominates the high latitude middle atmosphere, which can be attributed to the annual cycle of solar insolation. The maximum amplitude of the annual cycle is observed at 80~ or S in the temperature data and a marked hemispheric di fference is observed II. Large longitudinal assymetry has also been established in AO over the tropicS l2

.

Even though the amplitude of AO is comparatively small at' low latitudes, it is found to be quite significant in both zonal and meridional winds. Figure 2 shows the height variation of the amplitude of AO and QBO in zonal wind and AO in meridiona l wi nd . Maximum amplitudes of the order of 19 mls and 3 mls are observed between 12 and 14 km in zonal and meridional winds, respectively. Average phase variation of AO and QBO with height are depicted in Fig. 3. The amplitudes of AO in the zonal wind obtained by Nagpal et at. 13 for Thumba, Shar and Ralasore were 18 mis, 22 m/s and 26 mis, respectively _ Reddy et aZ. 4 also observed si mi lar

157

128

92

46

171 167

143 124

73 72

41 34

20

~ 15 iii 0

E10 -' D.

~ 5

135

112

65

24

58

58

58

55

3 4 5 6 FREQUENCY,cycIeaIyr

12km

7 8 9

10.---r---r---r---r---~~---r--~--,

29km

7 8 9

Fig. I-FFT spectra for zonal winds at different heights for the ent ire time series during 1977-1 993 over J-lyderabad

35

30

25

~ iii

~2O oJ.

15

10

5 -10

ZONAl

l-<lBO

2-1'.0

0 10 20 AMPUl1JDE,mll

30

MERIDIONAl

30

25

15 AO

10

~L---~O~--~5~--~10 AMPuruoE. trW

Fig. 2-Height profi les of amplitudes of AO, QBO in zonal wind and AO in meridional wind fo r the period 1977- ) 993 over Hyderabad derived by usi ng harmonic analys is­least square fi tting technique

results . The 0 phase (time of max imum westerly wi nd) is fo und to decrease slowly with height as shown in Fig. 3, which is comparab le to those observed by others .

228 INDIAN J RADIO & SPACE PHYS, ocrOBER 1999

35 35

oeo

30 30

25

\ 25

i i

120

----' \--uI

~20 :i. - ,.- :i.

\ I

15 15

- I-

10 10

-L

~ 0 2 .. SO 0 20 40 PHASE, mcnIh PHASE, mcnIh

Fig. 3---Height profiles of the phases of AO and Q80 in zonal wind fo r the period 1977- 1993 over Hyderabad derived by using harmonic analys is-least square fitting technique

3.2 Quasi-biennial oscillation

No evidence exists for truely periodic atmospheric oscillations apart from diurnal and annual components and their harmonics. The oscillation which comes closest to periodic behaviour is the QBO in the mean zonal wind of equatorial stratosphere, with a period varying from about 22 to 30 months. Vertically propagating equatorial Kelvin and Rossby-gravity waves provide the zonal momentum sources necessary to drive QBO. Holton and Lindzen 14 have successfully explained different observed features of the QBO.

In recent years, QBO in the zonal wind has gained a lot of importance and is considered to control both lower and upper atmospheric dynamics and chemistry. Kane l confirmed the presence of QBO in ionospheric parameters such as foE, foF2, hmF2, etc. Recently Saji Abraham et al. 16 investigated long-term variations in ionospheric absorption data over Delhi, and found a remarkable anti-correlation between the

QBO in wind data at 28 km (Thumba) and absorption data at the mesospheric level. There are also several studies relating the lower stratospheric QBO and EL Nino events I7

-20

.

Results of the present study of amplitude and phase varIatIOns of this important oscillation over Hyderabad are shown in Figs 2 and 3. Horizontal bars show the standard deviations of wave amp litude and phase. It is observed that the QBO amplitude starts becoming prominent above 18 km and maximizes at 29 km '(9 mls) with a period of 29 months. The QBO amplitude is found to decrease systematically w ith

latitude. Nagpal et al. 13 found the amplitudes of QBO to be 20 mis, 10 mls and 4 mls for the stations Tl1umba, Shar and Balasore, respectively. The heights of peak QBO amplitude as observed by them were 28 km, 33 km and 37 km, respectively. Sasi el al6 have shown that QBO in the zonal wind at Trivandrum (8.5"N) is well defined and cycles are easi ly identifiable, whereas it is not so well pronounced for an extra tropical station like Balasore (25.1"N). Figure 4 shows a time-series of zonal wind veloc ities at different heights. Thirteen-point running average has been taken to smooth the data and to filter out short period oscillations. Downward phase propagat ion can be seen from the height variation of westerly maxima .

3.3 Semi-annu al oscillation and terr-a nnual oscillation

The amplitudes of SAO and TAO in the troposphere and lower and middle stratosphere are found to be quite insignificant as reported by earlier workers.4

,13 and hence not been considered for further analysis.

3.4 Solar activity dependence of wind components

The solar activity dependence of AO, QBO and prevail ing wind has been studied using the method described earlier. The results of harmonic analysis show higher AO amplitude in the zonal wind during high activity periods HA I and HA2 having amplitudes of the order of 25 mls and 26 mis, respectively, compared to its va lue during LA (15 m/s). The height of maximum AO amplitude is also found to vary with so lar activity havi ng altitudes of 14 km and 13 km during HA I and HA2 and as low as 10 km during the LA period . The AO component in meridional wind did not show any regular variation with solar activity.

Figure 5 shows the height variation of the AO and QBO amplitudes for the different periods HA I , HA2 and LA. The amplitude. of QBO also shows appreciable variability with solar activity particularly at higher altitudes. The period of QBO is found to be longer in high activity periods, being around 30 and 29 months, compared to 22 months during the low activity period. The corresponding height of the QBO amplitude peak is 29, and 30 km for HA I and HA2 and 33 km for the LA period . Sasi et al. 6 analyzed wind data over Trivandrum and showed that the standard deviation of maximum easter ly amplitude was more than that of the westerly maximum, indicating greater variation from cycle to cycle. The/ found that the mean period of the westerly regime was longer than that of easterly regime except at altitudes

)

•

•

ALEEM BASHA et al. : ATMOSPHERIC OSCILLATIONS IN TROPOSPHERE & STRATOSPHERE 229

100 .. 120 1«1 160 180 200 TIME, mcnIh

Fig. 4-Tcmporal vari ation of zonal winds at 4 heights betwt'~n

10 and 25 km (Successive plots are displaced by 20 mls)

35 35

HA2 HAl LA HAl - I' I -",,'

30 , I 30 LA."· ~ HAZ , , \

, :'

25 25 I

/

! ! ,

ui ui g20 g20

I ,

t:: t:: .... ~ \ I

...I ... ... 15 15

, /

10 10

6 0 10 20 30

6 0 6 10 15

AO AMPLITUDE, mi. aBO AMPLITUDE, mi.

Fig. 5--\-Icight protiles of the amp litudes o f the AO and QBO for th ree so lar acti vity periods. IIAI(1978-83). II A2 ( 1988-93) and LA ( 1983-87)

from 15 to 18 km where the easterly period was longer. Similar analysis for the station, Balasore, showed an increasing trend of QBO amplitude with height. The mean periods at Balasore were nearly the same as those ofTrivandrum in the altitude range 25 -34 km, whereas above 34 km and below 25 km the easterly and westerly periods were shorter than 12 months in contrast to the behaviour at Trivandrum where they remained longer than 12 months in the altitude range 15 - 42 km.

To study the cross-correlation coeffic ients between the so lar parameters and AO. harmonic analysis has been carried out for individual year's data . The amplitudes of AO, so obtained, have been averaged over a slab of 8 km (between 10 and 17 km) which is the e-folding depth around the he ight o f maximum AO amp litude. T he cross-correlation coefficients (r) between the AO amp litudes and sunspot numbers (RJ and F lo7 cm solar nux are 0 .78 and 0.75 for zona l wi nd and 0.26 and 0 .20 fo r meridional wind , respectively. To consider the tatisti cal significar ce of

these correlation coefficients, I-tests have been carried out. A I-test for the significance of r is given by

t=r~(n-2) / ~(l-r 2 ) ... (I)

which has n-2 degrees of freedom. The I values obtained for the zonal wind case .are 4.44 and 4.80 and those for the meridional wind are 0 .80 and 1.05. The critical value for I with 15 degrees of freedom and at 0.01% level of s ignificance IS 2.947. Because the test statistics (zonal wind AO) fall into the upper critical region, it is concluded that there is a rea l correlation between the amplitude of AO in the zonal wind and Rz and F IO .7 cm solar flux, whereas the correlation of the amplitude of AO in meridional wind and the solar parameters are not significant. Figure 6 shows the AO amplitudes for each year (both zonal , and meridional), for the 8 km height-s lab average, between 1977 and 1993 with the corresponding values of sunspot number and F IO.7 cm so lar flu x .

To study the correlation of prevailing wind with so lar parameters, two hei ght regions (average 10-15 km and 25-30 km) have been considered. The va lues of cross-correlation coefficients of zonal and meridional prevailing winds with Rz for the first and second s labs are 0.21, - 0. 18 and 0.26 , - 0 .29 respectively. Results of I - tests carried out in a similar manner as, mentioned earlier, show that the correlations are not s ignificant. Greisiger el a/9 , Babajanov el al. 21 and Namboothiri el al. lo studied solar activity variations of wind parameters for upper atmosphere over mid-latitude stations. Their results are contradictory in nature. S imilar studies for lower atmosphere and low latitude stations are not available fo r comparison with the present work.

3.5 Energy density of AO with altitude

T he energy density of AO has been calcu lated using the average dens ity (p) from the atmospheric model g iven by Sasi and Sen Gupta22 for Indian stations lying between 0° and 15°N following the standard relation

E = (1 /2) P V2

and

... (2)

where. V, and Vm are the mean values of zonal and meridional amplitudes of AO obtained by harmonic ,analys is. Figure 7 shows the a ltitude variation of energy density of AO over /-Iyde rabad. T he energy density is fo und to increase rap id l ti ll about 10 km

230 INDIAN J RADIO & SPACE PHYS, OCTOBER 1999

,-, :;] 88 so

Fig. 6-Variati on of the yearly AO ampli tude between ( 10-17) km in zonal and meridional 'winds with corresponding valucs of sunspot numbers (R,) and FlO 7 cm so lar tlux

and then to decrease sharply. Most of the energy appears to be confined in the troposphere be low the tropopause level. There is an indication that very little energy is leaked through the tropopause and gets di ssipated thereafter.

4 Summary and conclusions

The results obtai ned in the present study of osc illat ions and the so lar activity dependence of wind parameters using 17 years (1977-1993) wind data in the altitude range of 7 - 32 km for a tropical station can be summarized as follows:

( i) Strong AO component is observed in both the zonal and meridional winds at an altitude between 12 and 14 km, being of the order of 19 m/s and 3 mis, respectively.

( ii) The AO is found to become stronger with solar activity. The amplitudes of AO in the zonal wind shows a strong posit ive corre lation with the F I07 cm solar flux and Rz numbers. The he ight of maxi mum AO amplitude is also found to vary with solar

activity, being higher for the high activity periods and lower for the low activ ity period. The meridional wi nd shows poor correlation '.'l ith solar activity.

(iii) The QBO ampl itude in the zonal wind maximizes at 29 km. The amplitude and period of QBO are observed to be 9 mls and 29 months, respectively. The amplitude of QBO is not observed to get affected by the variation in solar activity. The period of QBO is found to be longer in hi gh activity periods, whereas the height of maximum QBO is more in the low activity period.

~,-----~----.------r-----~----~----~

30

1? 25

ui o I-

~20 5 «

15

10

5'~----7-----~----~----~----~----~ -1 0 2 3

LOG OF ENERGY DENSITY (AO). Jouleslml

Fig. 7- Altitude variation of the energy density of AO over I-Iyderabad

(iv) No clear trend of solar radiation dependence of the preva iling wind was observed in the present study. T his may be due to the short length of the data set. Data from a number of tropical stations over more so lar cycles should be studied to confirm the relation of the prevailing w ind with so lar activ ity.

Acknowledgements

The authors express the ir sincere thanks to Prof. S V Dam Ie, Chairman, Ballooll Facili ty of T IFR and to India Meteorolog ical Department, Pune, for providing the data used in this study. Three of the authors (GD, HAB and BB) express their si ncere thanks to Nawab Shah Alam Khan, Chairman and Mr.Mahboob Alam Khan, Secretary of A.U .College for their active support. The authors also wish to thank the anonymous referees for their val uable com ments which have substantia lly improved the qua lity of th is paper.

References I Reed R J, J Atmos Sci (USA) , 22 ( 1965) 33 1. 2 Reed R J, J Geophys Res (USA) , 17 (1966) 4223. 3 Belmont A D, Dartt 0 G & Nastrom G 0, Q J R Meteorol

Soc (UK) , 100 (1 974) 203. 4 Reddy C A, Raghava Reddi C & Mohankumar K G. Q J R

Meteoro/ Soc (UA) , 112 ( 1986) 8 11 . 5 Plamb R A, Dynamics of the Middle A tmosphere, edited by J

R I-I olton and T Matsuno, (Terra Sc ientific Pu bli shing Corporat ion, Tokyo), 1984, p 273 . ~

5 Sasi M N & Murthy B V K, JAIm es & Terr Phys (UK). 53 ( 199 1) 11 73 .

7 Chakravarty S C, Jayati Datta & Revankar C I), Cllrr Sci (India) 63 (1992) 33.

8 Sprenger K & Schmindcr R, J ;l Imos & Terr Phys (UA). 3 1 ( 1969) 217 .

ALEEM BASHA et al.: ATMOSPHERIC OSCILLATIONS IN TROPOSPHERE & STRATOSPHERE 231

9 Greisiger K M, Schminder R & Kurschner D, J A lmas &Terr Phys (UA.), 49 (1987) 281.

1.0 Namboothiri S P, Manson A H & Meek C E, J Almas &Terr Phys (UK), 55 (1993 ) 1325.

11 Barnett J J & Corney M, Handbook for MAP, USA. , Vo1.l6, 1985, p. 86.

12 Reddy C A & Raghava Reddi C, J Almas &Terr Phys (UA.), 48 (1986) 1085 .

13 Nagpal 0 P, Jain A R & Dhaka S K, Indian J Radio & Space Phys, 18 (1989) 233:

14 Holton J R & Lindzon R S, J Almcs Sci (USA), 29 (1972) 1076.

15 Kane R P, J Almas & Terr Phys (UK), 57 ( 1995) 41 5. 16 Saji Abraham, Dhaka S K, Praveen K D, Nath N, Nagpal 0 P

& Baluja K L, J A lmas & Terr Phys (UK) , 59 (1997) 1885. 17 Yasunari T, J Mel Soc Jpn (Japan) , 67 ( 1989) 483. 18 Xu 1, J Almos Sci (USA), 49 (\992) 725 . 19 Angell J K, Geophys Res Lett (USA), 19 (1992) 285. 20 Sasi M N, J Atmos & Terr Phys (USA ), 56 (1994) 1563 . 21 Babajanov P B, Kolmakov V M & Chebotarev R p, Aslr Circ,

No.950, 1977. 22 Sasi M N & Sen Gupta K, Scientific Report SPL SR:006:85,

Space Physics Laboratory, Vikram Sarabhai Space Centre, Trivandrum, India, 1986.